Re: [Dime] RFC 3588 (Diameter Base Protocol) - Section 2.1.1. SCTP Guidelines

[Sebastien Decugis <[e-mail]>]

 Hi,

> How is an end-point sending a CEA and then sending a request over separate SCTP streams meant to work-around this issue? 
Sorry if it was not clear, I meant it as "the only exception to
out-of-order delivery *not* being an issue".
So, what I was describing is an actual issue (in my opinion) with
multi-stream use.

It can be worked around in the implementation, for example by allowing
messages to be sent over a different stream only after a new message has
been received from the remote peer (confirming the remote peer moved the
connection to OPEN state), or after a reasonable delay (for example a
few seconds). If the first option is chosen, it can be a good practice
to always send a DWR after the CEA is received (as it is done after a
network failure recovery).

I hope this clarifies my point ^^.

Sebastien.

-- 
Sebastien Decugis
Research fellow
Network Architecture Group
NICT (nict.go.jp)